---
id: "concept-ai-knowledge-hiding"
type: "concept"
source_timestamps: ["¶3", "§ What Our Research Shows"]
tags: ["shadow-ai", "organizational-behavior", "core-thesis"]
related: ["concept-suppression-of-solutions", "claim-trust-predicts-hiding", "framework-costs-of-ai-visibility"]
definition: "The intentional concealment of AI tools, prompts, and workflows from employers and colleagues, driven primarily by a lack of organizational trust."
sources: ["execution"]
sourceVaultSlug: "hbr-seg-execution"
originDay: 8
articleStem: "hbr-cl-76-employees-not-transparent-ai-usage"
sourceUrl: "https://hbr.org/2026/06/why-employees-arent-transparent-about-their-ai-usage"
sourceTitle: "Why Employees Aren’t Transparent About Their AI Usage"
---
# AI Knowledge Hiding

The intentional concealment of AI use — prompt sequences, chained tools, and successful iterations — from employers and peers. This is the phenomenon popularly called **'shadow AI.'**

**Prevalence is high:** 57% in the [[entity-kpmg-melbourne-study|KPMG / University of Melbourne study]]; 30.3% in the authors' own survey of daily AI users. Critically, it is **not** driven by ignorance of collective value — nearly **80%** of employees know that sharing helps the team. It is driven by a lack of organizational trust, fear of replacement, and the desire to maintain a competitive edge.

The deepest predictor is trust itself — see [[claim-trust-predicts-hiding]]. The employee's decision calculus is decomposed in the [[framework-costs-of-ai-visibility]], and the collective-solution variant is [[concept-suppression-of-solutions]]. The economic cost is summarized in [[quote-roi-kept-by-employee]].

**Enrichment / lineage:** The foundational construct is Connelly et al. (2012), which defines knowledge hiding as intentional concealment of *requested* knowledge (via evasive hiding, playing dumb, or rationalized hiding). Newer studies (Frontiers; Nature) show AI awareness and perceived replaceability are positively correlated with hiding, partially mediated by psychological availability — supporting, but also complicating, the article's trust-centric framing.


## Related across articles
- [[claim-policing-ai-impossible]]
- [[concept-workslop-d8]]
- [[claim-genai-not-displacing]]


## Related across segments
- [[concept-shadow-ai]]
- [[concept-clandestine-ai-use]]
- [[concept-efficiency-tax]]